This endpoint allows similarity comparisons on existing or new products against an existing catalog.
The input is a set of products, which can either be existing products in the catalog (item_id) or new products (items).
For each item, the output is a ranked list of the most similar products in the catalog from rank 1 to num_similar_products. Rank 1 reflects the most similar product.
By default, similarity takes into account both product images and metadata.